Green CovR 2020-2021 data report

Data Collection

In the fall of 2020 and spring of 2021 184 photos of cover crops were submited from 7 counties. Photos were taken with a variety of devices, looking straight down from about 5 ft off the ground. In addition, Soil and Water Conservation District employees uploaded information about cover crop planting date, method, species, and previous crop.

The primary analysis we performed was to estimate the green ground cover in each photo, as a metric of cover crop success. A program called Green CovR (https://fgcc-app.github.io/) was used to analyze the % of pixels that were green, i.e. filled by a living cover crop. As seen in the table below, a wide range of cover crop success was was seen across the state. The highest mean and median were seen in Winona County, and lowest in Pipestone.

Table 1: % Green Cover by County

county	Minimum	Maximum	Average	Median	Data.Points
Becker	0	95	33	37	21
East Otter Tail	0	99	22	15	40
Faribault	3	86	30	22	59
Mower	0	94	39	28	22
Olmsted	4	95	44	41	13
Pipestone	1	16	7	3	5
Traverse	2	98	26	14	15
Winona	9	91	57	66	9

As Photo 1 demonstrates, some of the low numbers may be due to interference with residue. This method is not a reliable estimate of soil protection for this reason. The % green cover must only be interpreted as a relative indicator of cover crop success. Some research is underway to test factors such as light angle on the ability of the software to detect living cover crops, and to correlate the Green CovR output with actual cover crop biomass. The photo showing 65% green cover also shows that you don't need to get anywhere near 100% to have highly effective ground cover.

Photo 1: Oat-radish-turnip mixture, 8% coverage and Oat-field pea-radish mixture, 64% coverage

Research Questions

The primary question we're able to address with this data is what is the relationship between cover crop success and planting date? Figure 1 shows a noisy decline with later planting.

There are a number of potential complicating factors: location, cash crop rotation, cover crop species and winter hardiness, and whether cover crops were interseeded, and we will explore these to the best of our ability.

It would be expected that cover success would increase from north to south. This is not broadly true. Considering the median to reduce the impact of outliers, the three southern counties represented, Pipestone, Olmsted, Winona and Faribault, have a wide range of cover, and two of these counties have such low numbers of participants that we can't draw robust conclusions from them. Of the three northern counties, Becker had substantially greater median green cover than East Otter Tail or Traverse. This may be due to the fact that both East Otter Tail and Traverse SWCDs took data from farmers participating in cover crop cost-share programs beginning in 2019, so they likely included many farmers new to managing cover crops.

Effect of interseeding

Figure 2 compares interseeding versus non in each location. The strong relationship between planting date and coverage holds only when cover crops were not interseeded. When cover crops were interseeded, there's not much increase in ground cover with earlier seeding dates. Interseeding cover crops are mostly not limited by growing degree days. Instead, moisture and light competition with the cash crop were likely the primary factors affecting interseeded cover crop success. Comparing interseeding vs not diretly, only in Traverse was there a significant disadvantage to interseeding.

Cash Crop Rotation

The crop a cover crop follows is a primary determinant of cover crop planting date, which in turn usually dictates cover crop species chosen. Here, interseeded and non-interseeded cover crops will be shown separately. Without enough data for robust comparison, it apears that none of the particular prior crops used were better suited for interseeding than others. The data does show that farmers across the state are experimenting with very early interseeding in field corn. For non-interseeded cover crops, greater coverage was generated when cover crops followed corn silage, and canning crops. FOr the more common cash crops, we saw fewer high values in corn grain (almost all <50% coverage), while soybean and small grain had greater variability. Seeding a cover crop after dry beans in East Otter Tail county had similar success to corn grain before Sept 14.

Effect of cover crop species

The question of cover crop species is a bit more complicated. We collected up to 4 species included in mixes, but did not include estimates of what the proportion of each species was seeded or emerged. So there's a limit to how much we can say about the effect of species here.

One major split among species is whether they were winter-killed or over wintering. In Figure 4, whether interseeded or not, winterkilled cover crops success declined more dramatically with later planting dates. Overwintering cover crops had a chance to catch up in the spring, so later planting dates could still produce a reasonable amount of cover. However, there is still a decline in cover produced with later planting, and no few fields planted after October 15 achieved over 50% green cover.

What to plant late in the year is of critical interest to Minnesota farmers, as most row crop producers don't have equipment to interseed, or would prefer not to potentially compete with a cash crop. The Midwest Cover Crop Council Decision Tool recommends planting winter rye up to Nov 20th, and winter triticale, winter wheat, winter barley or camelina until October 25 (averaged across Minnesota). In our dataset, we captured 95 stands of cereal rye and 15 fields with winter triticale, so Figure 5 focuses on these known and widely-used winter-hardy species. Cereal rye's good reputation as a reliable choice for late planting dates is supported by the very moderate decline in green cover with planting date. Cereal rye is not a likely candidate for early interseeding, as green cover increased with later-planted interseeded crops (probably corresponding to reproductive-stage cash crops). In many of these interseeded cases, cereal rye was part of a larger mix, and so the total cover can't be attributed just to that species.

Because of the widespread use of mixes, it's difficult to isolate effects of any one species. When a single species was used (51 photos), it was most often cereal rye (41 photos), with a few occurances each of oats, winter triticale, and winter wheat. Mixes usually contained a grass and a brassica, and some also contained a legume. Below can be seen the most popular species used.

Species	Number.of.Uses
CEREAL_RYE	96
TURNIP	52
RADISH	51
OATS	47
ANNUAL_RYE	20
WINTER TRITICALE	16
RED_CLOVER	13
FIELD_PEA	12
CRIMSON_CLOVER	12
SPRING_WHEAT	11
RAPESEED	9
COWPEAS	6
HAIRY_VETCH	3

Next Steps

Data collection will be ongoing in fall 2021 and spring 2022 with SWCD staff partners. The dataset will be presented in presentations and online statewide, and each partner can use it for local meetings as they see fit. The dataset may also be used in conjunction with further method development for the Green CovR software (a collaboration with Melissa Wilson and Manuel Sabbbagh).